Re: [Ganglia-general] Ganglia 4.x architecture planning
On Mar 27, 2014, at 3:07 PM, Daniel Pocock dan...@pocock.com.au wrote: The introduction of RabbitMQ is an optional dependency. It would allow users to send commands from the web interface. 1. Why add the extra dependency on rabbitmq? As long as you're adding a persistent data store (mongo), couldn't it also serve as a message queue? 2. Why will the web UI start using Python flask? alex -- ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Why IP instead of hostname?
On Jun 28, 2013, at 3:56 AM, Benjamin Wang (gendwang) gendw...@cisco.com wrote: Hi, Now I deploy two gmond in two hosts(10.74.125.75 and 10.74.125.85). If I change the /etc/ganglia/gmetad.conf as following, then 10.74.125.85 will be showed by the hostname, but the 10.74.125.75 will be showed by IP data_source Design Team 10.74.125.85 10.74.125.75 But if I change the /etc/ganglia/gmetad.conf as reversed, then 10.74.125.75 will be showed by the hostname, but the 10.74.125.85 will be showed by IP data_source Design Team 10.74.125.75 10.74.125.85 What is the issue? Try adding both hosts to the /etc/hosts file on the machines running gmond, then restart gmond. alex -- This SF.net email is sponsored by Windows: Build for Windows Store. http://p.sf.net/sfu/windows-dev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] How to query local gmond for local data
On Oct 22, 2012, at 5:18 AM, Bostjan Skufca wrote: Hi, there is a way to query gmond headnode for any data, I use it, it works well. My question is: is there a way to query local gmond instance for local data, without it being a headnone? Further explanation of context: Currently data is being collected by gmetad, which is connecting to headnodes to retreive it. Non-headnodes use unicast to send data to headnodes. Works quite fine. On top of this, there is monitoring with nagios which connects to gmetad and queries it for various data. There are two drawbacks to this mode: - as number of nodes increase, load increases on gmetad and monitoring host, linearly - there are three service dependencies, local gmond, headnode gmond and gmetad. If any of these fail, metric ceases to be monitored and results in warning. This dependency setup is a nightmare to manage, especially if you have to manage more than one cluster of servers, which has its own headnodes and its own gmetad server. Hi Bostjan. Have you considered adding multiple head nodes to each unicast cluster? This would remove the single points of failure you describe. 1. Each node in the cluster reports its metrics to 2 (or more) nodes. 2. gmetad.conf lists all of these head nodes in the data_source line. If the first gmond stops answering, gmetad will try the next one w/ 0 service interruption. I would like to replace current setup with NRPE call + local gmond query ON the machine being monitored, to shift the work to the machine being monitored (except for TCP connection to NRPE, of course, but this is lightweight compared to previous solution). This also eases the dependency management in nagios, because the only dependencies are local to the monitored machine - nrpe and gmond. Thank you, b. -- Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_sfd2d_oct ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Custom JSON graphs
Future reference: http://jsonlint.com/ is handy for validating JSON.
On Oct 4, 2012, at 10:01 AM, Ron Cavallo wrote:
Never mind I figured it out. Missing comma. Thanks Jochen.
Ron Cavallo
Sr. Director, Infrastructure
Saks Fifth Avenue / Saks Direct
12 East 49th Street
New York, NY 10017
212-451-3807 (O)
212-940-5079 (fax)
646-315-0119(C)
www.saks.com
-Original Message-
From: Ron Cavallo [mailto:ron_cava...@s5a.com]
Sent: Thursday, October 04, 2012 10:49 AM
To: Jochen Hein; ganglia-general@lists.sourceforge.net
Subject: Re: [Ganglia-general] Custom JSON graphs
Have you been able to graph custom metrics via json reports? I tried
everything. I even tried using the below to see if the problem was my
metric. load_one is graphed by the cpu json report, so I KNOW ganglia
can use it. Still my Web Report shows up No Matching Metrics Detected
{
title : Web Servers,
vertical_label : Seconds,
report_name : webs_report,
report_type : standard,
series : [
{ metric: load_one, color: bb, label:
Web1, line_width: 2, type: line },
]
}
Ron Cavallo
Sr. Director, Infrastructure
Saks Fifth Avenue / Saks Direct
12 East 49th Street
New York, NY 10017
212-451-3807 (O)
212-940-5079 (fax)
646-315-0119(C)
www.saks.com
-Original Message-
From: Jochen Hein [mailto:joc...@jochen.org]
Sent: Thursday, October 04, 2012 10:44 AM
To: Ron Cavallo
Cc: ganglia-general@lists.sourceforge.net
Subject: Re: [Ganglia-general] Custom JSON graphs
Ron Cavallo ron_cava...@s5a.com writes:
Now I have this problem:
In ganglia 3.4.2, I am trying to get the following to work:
- I created a webs_report.json file in graph.d.
- I named a custom metric that I am collecting via gmetric from the
host named in the json, like this:
{
report_name : webs_report,
report_type : standard,
title : Web Server Response Time Report,
vertical_label : Seconds,
series : [
{ metric: Ext-NewWeb01-Entry, color: bb, label: Web
1, line_width: 2, type: line }, { hostname:
SD1GGL01LX.SAKSDIRECT.COM clustername: SaksGanglia
^here is a comma missing
},
That comma might be the problem.
]
}
Hope that helps
Jochen
--
The only problem with troubleshooting is that the trouble shoots back.
--
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly what is
happening inside your Ruby, Python, PHP, Java, and .NET app Try New
Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
--
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and .NET app
Try New Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
--
Don't let slow site performance ruin your business. Deploy New Relic APM
Deploy New Relic app performance management and know exactly
what is happening inside your Ruby, Python, PHP, Java, and .NET app
Try New Relic at no cost today and get our sweet Data Nerd shirt too!
http://p.sf.net/sfu/newrelic-dev2dev
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] [Ganglia-developers] Ganglia not recognizing other nodes?
Good summary. Another way to say it: it's gmond, not gmetad, which decides
which hosts are members of the cluster.
On Jul 21, 2012, at 11:18 AM, Jonah Horowitz wrote:
Jeff,
Gmetad doesn't exactly treat the nodes in the data_source line the way you're
thinking it does. Gmond assumes all hosts have a full set of data and only
uses the second one if the first can't be contacted. If you want the two
nodes to be in the same cluster, you have to configure the gmond on each host
to join the same multicast group (or setup unicast).
On both nodes, you should have a section of your gmond that looks like this:
udp_send_channel {
mcast_join = 239.2.11.71
port = 9649
ttl = 64
}
udp_recv_channel {
mcast_join = 239.2.11.71
port = 9649
bind = 239.2.11.71
}
Make sure your switches support multicast. If not, use unicast:
#on both hosts
udp_send_channel {
host = 10.1.0.250
port = 9649
}
#on master node
udp_recv_channel {
port = 9649
}
send_metadata_interval 30
If you're going to scale, unicast works better anyway.
Hope that helps,
Jonah
On Jul 21, 2012, at 8:13 AM, Jeff Layton wrote:
Jesse,
[Note - I changed groups to ganglia-general since this isn't
a developers issue - just a silly user issue].
I must admit that networking is one of my weak areas but here
is are some relevant sections of output from netstat and lsof
on the master node.
Netstat:
[root@test1 ~]# netstat | more
Active Internet connections (w/o servers)
Proto Recv-Q Send-Q Local Address Foreign Address
State
tcp0 0 192.168.1.250:53687 ord08s06-in-f15.1e100:https
ESTABLISHED
tcp0 0 10.1.0.250:shelln0001:48199
ESTABLISHED
tcp0 0 192.168.1.250:41461 ord08s08-in-f21.1e100:https
ESTABLISHED
tcp0 0 192.168.1.250:41476 ord08s08-in-f21.1e100:https
ESTABLISHED
tcp0 0 10.1.0.250:8649 10.1.0.250:49899
TIME_WAIT
tcp0 0 10.1.0.250:nfs n0001:imaps
ESTABLISHED
tcp0 0 192.168.1.250:50191 den03s05-in-f16.1e100:https
TIME_WAIT
tcp0 0 10.1.0.250:8649 10.1.0.250:49905
TIME_WAIT
tcp0 0 192.168.1.250:52133 ord08s07-in-f14.1e100:https
ESTABLISHED
tcp0 0 192.168.1.250:50500 ord08s09-in-f22.1e100:https
ESTABLISHED
tcp0 0 10.1.0.250:8649 10.1.0.250:49904
TIME_WAIT
tcp0 0 10.1.0.250:8649 10.1.0.250:49903
TIME_WAIT
tcp0 0 10.1.0.250:43479n0001:ssh
ESTABLISHED
tcp0 0 192.168.1.250:52134 ord08s07-in-f14.1e100:https
TIME_WAIT
tcp0 0 192.168.1.250:53686 ord08s06-in-f15.1e100:https
ESTABLISHED
udp0 0 192.168.1.250:52035 239.2.11.71:8649
ESTABLISHED
lsof output:
[root@test1 ~]# lsof | grep gmond
gmond 1948 nobody cwd DIR8,2 4096
2 /
gmond 1948 nobody rtd DIR8,2 4096
2 /
gmond 1948 nobody txt REG8,2111475
2527491 /usr/sbin/gmond
gmond 1948 nobody mem REG8,2161084
1318670 /lib/libexpat.so.1.5.2
gmond 1948 nobody mem REG8,2 43916
2527474 /usr/lib/libconfuse.so.0.0.0
gmond 1948 nobody mem REG8,2 67920
3165062 /usr/lib/ganglia/modcpu.so
gmond 1948 nobody mem REG8,2 67461
3165065 /usr/lib/ganglia/modload.so
gmond 1948 nobody mem REG8,2131044
1318655 /lib/libpthread-2.12.so
gmond 1948 nobody mem REG8,2 1876580
1318624 /lib/libc-2.12.so
gmond 1948 nobody mem REG8,2113908
1318641 /lib/libnsl-2.12.so
gmond 1948 nobody mem REG8,2 67515
3165068 /usr/lib/ganglia/modnet.so
gmond 1948 nobody mem REG8,2190604
1318668 /lib/libpcre.so.0.0.1
gmond 1948 nobody mem REG8,2142480
1318667 /lib/ld-2.12.so
gmond 1948 nobody mem REG8,2 67469
3165063 /usr/lib/ganglia/moddisk.so
gmond 1948 nobody mem REG8,2 58704
1318647 /lib/libnss_files-2.12.so
gmond 1948 nobody mem REG8,2 67885
3165066 /usr/lib/ganglia/modmem.so
gmond 1948 nobody mem REG8,2184012
2527486 /usr/lib/libapr-1.so.0.3.9
gmond 1948 nobody
Re: [Ganglia-general] ganglia-gweb 3.4.2 (Perzl-packages)
Since inspect_graph.php is making its own http request, it's not going to be aware of the auth credentials you've supplied from your client. Hence file_get_contents() gets a 401 response from the server. It might help to alter your authentication configuration to allow all requests from localhost. (Just thinking out loud.) alex On May 22, 2012, at 8:20 AM, Jochen Hein wrote: Hello, I've upgrade our ganglia server to ganglia-gweb 3.4.2. We use external authentification to ganglia. In httpd.conf we have: Directory /var/www/htdocs/ganglia AuthType Basic AuthName Ganglia AuthUserFile /opt/freeware/etc/httpd/http_user Require valid-user /Directory When I first open the Web frontend, I have to enter username and password and can then use Ganglia. When I try to inspect a graph, I get in the error_log: [Tue May 22 14:23:43 2012] [error] [client x.y.z.7] PHP Warning: file_get_contents(http://ganglia.example.com/ganglia/graph.php ?flot=1amp;h=myhostamp;m=cpu_usedamp;r=houramp;s=by%20nameamp;hc=4amp;mc=2amp;st=1337689406amp;g=cpu_used_reportamp;z=la rgeamp;c=B3) [a href='function.file-get-contents'function.file-get-contents/a]: failed to open stream: HTTP request failed! HT TP/1.1 401 Authorization Required\r\n in /var/www/htdocs/ganglia/inspect_graph.php on line 28, referer: http://ganglia.example.com/g anglia/?r=hourcs=ce=m=cpu_useds=by+namec=B3h=myhosthost_regex=max_graphs=0tab=mvn=sh=1z=smallhc=4 [Tue May 22 14:23:43 2012] [error] [client x.y.z.7] PHP Warning: Invalid argument supplied for foreach() in /var/www/htdocs/ga nglia/inspect_graph.php on line 30, referer: http://ganglia.example.com/ganglia/?r=hourcs=ce=m=cpu_useds=by+namec=B3h=myhosthost_regex=max_graphs=0tab=mvn=sh=1z=smallhc=4 ERROR: Garbage found where STACK expected Any ideas how that can be fixed? Jochen -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Ganglia for Windows
On Mar 30, 2012, at 5:17 AM, Burton, Steven wrote: Bernhard, it doesn't seem likely that I'd be much help as I've come unstuck trying to install the web frontend. I'm beginning to suspect that there are two different web frontends and that the documentation for v3.3.1 is held in a secure secret location which while laudable in itself isn't really helping. I'm currently using the files from gweb-3.3.1.tar.gz and the instructions from http://sourceforge.net/apps/trac/ganglia/wiki/ganglia-web-2#Installation I previously tried the files from ganglia-3.3.1/web but that didn't lead very far. In both cases the problem seems to be setting the paths to various components. I've currently got it down to one error in the web interface, when running index.php: Fatal error: Call to undefined function xml_parser_create() in /usr/local/www/ganglia2/ganglia.php on line 339 Steve. xml_parser_create is part of the XML Parser extension. http://us.php.net/xml_parser_create This extension is enabled by default. It may be disabled by using the following option at compile time: --disable-xml http://us.php.net/manual/en/xml.installation.php Is it disabled in your PHP? Try looking at phpinfo() output to check. alex signature.asc Description: Message signed with OpenPGP using GPGMail -- This SF email is sponsosred by: Try Windows Azure free for 90 days Click Here http://p.sf.net/sfu/sfd2d-msazure___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Upgrade to 2.2.0 frontend
On Mar 7, 2012, at 9:39 AM, Ron Cavallo wrote: Copying and using functions.php and graph.php from the old version into the new version was needed (too many errors). functions.php and graph.php are supposed to be replaced with each new release, and you shouldn't mix/match them across gweb releases. New functions may have been added, and existing ones may have changed behaviors, in the new release. The notice-level errors in graph.php can probably be ignored. Not as sure about the in_array() errors in functions.php, but I'm guessing that continuing to use the functions.php from 2.1.8 is probably going to do more harm than good. Could you open github issues for these 2 problems, so we have a record that these issues need to be dealt with? thanks, alex signature.asc Description: Message signed with OpenPGP using GPGMail -- Virtualization Cloud Management Using Capacity Planning Cloud computing makes use of virtualization - but cloud computing also focuses on allowing computing to be delivered as a service. http://www.accelacomm.com/jaw/sfnl/114/51521223/___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] node addition
On Dec 21, 2011, at 9:55 AM, solarflow99 wrote: just have each cluster group use a different port in its udp_send channel directive. Then in the server, run multiple gmond -c gmond.xxx.conf daemons all listening on their cluster specific ports. gmetad uses servername:portnumber for each datasource. Its actually a cleaner way to do it then using 1 server per cluster group. Use caution. This single box with all the gmonds is now a single point of failure for all your clusters. Any problems with this box mean you'll lose metrics from all your clusters. It sounds like you're differentiating a normal gmond from a head gmond. This isn't necessary. Any gmond can be a head node. (You can use one of your normal compute-node gmonds as the head node.) You don't need to set up special collection-only nodes for gmetad to poll. In fact, I think it's better to configure several head nodes, so if one does go down gmetad can switch over to a secondary data source and you will still be able to get metrics from your cluster. To do this via unicast: - pick 2 (or more) head nodes. These can collect all the metrics you want. The only thing that makes them 'head' nodes is that you configure udp_recv_channel for them. - all nodes in the cluster have udp_send_channels for ALL head nodes. - gmetad has ip:port for all head nodes in its data_source config for the cluster. In this configuration, gmetad will connect to the first gmond in data_source. But if that node dies, gmetad will fail over to the second gmond in data_source. Since all cluster nodes are reporting to both of these gmonds, they both have all cluster metrics and you don't lose anything. Now no gmond is a single point of failure. Of course the more head nodes you configure the more network traffic you create, so balance that against your availability concerns. 2 head nodes seems like a great first-guess rule-of-thumb to me. best, alex -- Write once. Port to many. Get the SDK and tools to simplify cross-platform app development. Create new or port existing apps to sell to consumers worldwide. Explore the Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join http://p.sf.net/sfu/intel-appdev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Adding hosts to cluster in gmetad
On your gmond machine, try 'telnet localhost 8649'. That will give you an XML
dump of all the metrics gmond knows about. I suspect you'll only see 1 host.
You may need a 'bind' IP in your udp_recv_channel configuration.
http://sourceforge.net/apps/trac/ganglia/wiki/Gmond%203.1.x%20General%20Configuration#udp_recv_channel
alex
On Dec 17, 2011, at 6:06 AM, Maciek Lasyk wrote:
Still haven't resolved the issues. Tiil now I haven't mentioned that
this is Ganglia 3.2.0 running on CentOs 5.6 and installed from sources
with standard ./configure --prefix=/usr/local/ganglia --with-gmetad
command
I changed some environment, and turned off gmond on gmetad host:
My gmetad.conf is now really simple:
=
debug_level 10
data_source RR1 192.168.0.6:8649 192.168.0.7:8649
gridname RTEST
rrd_rootdir /var/lib/ganglia/rrds_test
case_sensitive_hostnames 1
=
So now I'm trying to get data from only 2 hosts in the same network.
I;s configuration gmond.conf:
=
cluster {
name = RR1
}
udp_send_channel {
host = 192.168.0.6
port = 8649
ttl = 1
}
udp_recv_channel {
port = 8649
}
tcp_accept_channel {
port = 8649
}
=
After starting gmetad I see:
[root@thebe www]# service gmetad start
Starting GANGLIA gmetad: Going to run as user nobody
Sources are ...
Source: [RR1, step 15] has 2 sources
192.168.0.6
192.168.0.7
xml listening on port 8651
interactive xml listening on port 8652
cleanup thread has been started
Data thread 1153255744 is monitoring [RR1] data source
192.168.0.6
192.168.0.7
[RR1] is a 2.5 or later data stream
hash_create size = 1024
hash-size is 1031
hash_create size = 50
hash-size is 53
hash_create size = 50
hash-size is 53
Updating host mneme.somedomain.com, metric load_one
[...]
Updating host mneme.somedomain.com, metric swap_free
Writing Summary data for source RR1, metric disk_free
[...]
Writing Summary data for source RR1, metric part_max_used
[RR1] is a 2.5 or later data stream
Updating host mneme.somedomain.com, metric load_one
[...]
Updating host mneme.somedomain.com, metric swap_free
Writing Summary data for source RR1, metric disk_free
[...]
Writing Summary data for source RR1, metric part_max_used
[RR1] is a 2.5 or later data stream
and so on...
So nothing changed. Tried strace the problem:
mmap(NULL, 10489856, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_ANONYMOUS|MAP_32BIT, -1, 0) = 0x45a53000
mprotect(0x45a53000, 4096, PROT_NONE) = 0
clone(child_stack=0x46453200,
flags=CLONE_VM|CLONE_FS|CLONE_FILES|CLONE_SIGHAND|CLONE_THREAD|CLONE_SYSVSEM|CLONE_SETTLS|CLONE_PARENT_SETTID|CLONE_CHILD_CLEARTID,
parent_tidptr=0x464539d0, tls=0x46453940, child_tidptr=0x464539d0) =
17108
write(2, cleanup thread has been started, 31cleanup thread has been
started) = 31
write(2, \n, 1
) = 1
gettimeofday({1324122388, 853770}, NULL) = 0
gettimeofday({1324122388, 853889}, NULL) = 0
select(0, NULL, NULL, NULL, {7, 125353}Data thread 1168451904 is
monitoring [RR1] data source
192.168.0.6
192.168.0.7
[RR1] is a 2.5 or later data stream
hash_create size = 1024
hash-size is 1031
hash_create size = 50
hash-size is 53
hash_create size = 50
hash-size is 53
Updating host mneme.somedomain.com, metric load_one
So nothing here... tried with ltrace:
[root@thebe www]# ltrace -s 320 /usr/local/ganglia/sbin/gmetad -d 10
__libc_start_main(0x4029b0, 3, 0x7fff93d11078, 0x4082d0, 0x4082c0
unfinished ...
signal(13, 0x1)
= NULL
strlen(/usr/local/ganglia/etc/gmetad.conf)
= 34
malloc(35)
= 0x2049010
strcpy(0x2049010, /usr/local/ganglia/etc/gmetad.conf)
= 0x2049010
strlen(/usr/local/ganglia/etc/gmetad.conf)
= 34
malloc(35)
= 0x2049040
strcpy(0x2049040, /usr/local/ganglia/etc/gmetad.conf)
= 0x2049040
getopt_long(3, 0x7fff93d11078, hVc:d:p:, 0x60bde0, 0x7fff93d10ddc)
= 100
__strtol_internal(10, 0x7fff93d10cc8, 0)
= 10
strlen(10)
= 2
malloc(3)
= 0x2049070
strcpy(0x2049070, 10)
= 0x2049070
getopt_long(3, 0x7fff93d11078, hVc:d:p:, 0x60bde0, 0x7fff93d10ddc)
= -1
free(0x2049040)
Re: [Ganglia-general] multiple cluster directives
On Dec 15, 2011, at 9:45 AM, solarflow99 wrote: Is it possible to have multiple cluster directives in gmond? I am trying to have all different cluster nodes send to 1 gmond receiver in unicast. Can this be done, or do I have to have multiple gmond daemons running each with their own config file? I've always thought that each gmond must be in 1 and only 1 cluster. Someone please correct me if I've got that wrong. alex -- 10 Tips for Better Server Consolidation Server virtualization is being driven by many needs. But none more important than the need to reduce IT complexity while improving strategic productivity. Learn More! http://www.accelacomm.com/jaw/sdnl/114/51507609/ ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Gmetric Kilobyte unit representation
The graph is showing you a value of roughly 10.5 million Kilobytes, which matches what you sent via gmetric. The label 'Kilobytes' is just text, and not interpreted by ganglia in any way, so no unit conversion will be performed. If you want the graph to be in gigabytes, do the division in your script prior to calling gmetric. $ gmetric --conf=gmond.conf --name test_metric2 --value 10.57 --type float --units Gigabytes alex On Dec 12, 2011, at 2:01 PM, Aidan Wong wrote: Hi, I'm trying to pass a metric with Gmetric in Kilobytes but the unit representation in the graph is not correct. Example: $ gmetric --conf=gmond.conf --name test_metric2 --value 10570393 --type int32 --units Kilobytes For the command above, the graph represents 10570393 Kb as 10.57 Mb instead of 10.57 Gb. See picture attached. Can someone indicate how I should make the graph display the correct byte conversion? Also is it possible to search the archive mailing list items from the SourceForge forum list? http://sourceforge.net/mailarchive/forum.php?forum_name=ganglia-general Thanks graph.jpg-- Learn Windows Azure Live! Tuesday, Dec 13, 2011 Microsoft is holding a special Learn Windows Azure training event for developers. It will provide a great way to learn Windows Azure and what it provides. You can attend the event by watching it streamed LIVE online. Learn more at http://p.sf.net/sfu/ms-windowsazure___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general -- Learn Windows Azure Live! Tuesday, Dec 13, 2011 Microsoft is holding a special Learn Windows Azure training event for developers. It will provide a great way to learn Windows Azure and what it provides. You can attend the event by watching it streamed LIVE online. Learn more at http://p.sf.net/sfu/ms-windowsazure ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] help for authorization in ganglia
Michele: Hi. I thought we determined (off list) that your problem was due to
usage of magic_quotes_gpc. Extra \\ characters were being added to the cookie
value, and this broke unserialization. I committed a fix to handle this
situation, which should be part of the next gweb release. The most relevant
change is here:
https://github.com/vvuksan/ganglia-misc/commit/daa863919ffcba1620b454133887587917c3a641#L0L26
base64 encode/decode seems like another (and possibly better) way to handle
this situation, though.
alex
On Nov 18, 2011, at 2:55 AM, mcarpene wrote:
Hi Parthiban,
thank you very much.
I solved this stripping out parenthesis from the token array :)
Thank you so much for your answer.
This list has been very useful.
cheers,
Michele
Il giorno ven, 18/11/2011 alle 11.01 +0530, Parthiban P ha scritto:
Hi Michele,
I faced same issue and fixed by doing the following change in the
gweb-2.1.8 source
Error : Warning: array_keys() [function.array-keys]: The first argument
should be an array../lib/GangliaAuth.php
Issue : PHP Serialize()/Unserialize() issue.
Solution :
Quickfix
edit the file /lib/GangliaAuth.php
the following lines need to be modified,
line 26 $data = unserialize($_COOKIE['ganglia_auth']); --- $data =
unserialize(base64_decode($_COOKIE['ganglia_auth']));
line 71 setcookie('ganglia_auth', serialize( array('user'=$user,
'group'=$group, 'token'=$this-getAuthToken($user)) ) ); ---
setcookie('ganglia_auth', base64_encode(serialize( array('user'=$user,
'group'=$group, 'token'=$this-getAuthToken($user))) ) );
For more details.. about the PHP Serialize()/Unserialize() issue and the
solution
Please kindly refer
http://davidwalsh.name/php-serialize-unserialize-issues
With thanks,
Parthiban ( parthibanwa...@gmail.com )
On Fri, Nov 18, 2011 at 4:18 AM,
ganglia-general-requ...@lists.sourceforge.net wrote:
Send Ganglia-general mailing list submissions to
ganglia-general@lists.sourceforge.net
To subscribe or unsubscribe via the World Wide Web, visit
https://lists.sourceforge.net/lists/listinfo/ganglia-general
or, via email, send a message with subject or body 'help' to
ganglia-general-requ...@lists.sourceforge.net
You can reach the person managing the list at
ganglia-general-ow...@lists.sourceforge.net
When replying, please edit your Subject line so it is more
specific
than Re: Contents of Ganglia-general digest...
Today's Topics:
1. Re: [Ganglia-developers] Enumerating the current value of
metrics (eg load_one) on host graphics (Vladimir Vuksan)
2. help for authorization in ganglia (m.carpen)
3. Re: help for authorization in ganglia (Alex Dean)
4. Deleting nodes from the cluster database... (David Barr)
5. ganglia front-end configuration using filters (mcarpene)
6. Re: ganglia front-end configuration using filters
(mcarpene)
7. Re: gmond stops listening after 6 minutes (Lance Smith)
--
Message: 1
Date: Thu, 10 Nov 2011 11:51:46 -0500 (EST)
From: Vladimir Vuksan vli...@veus.hr
Subject: Re: [Ganglia-general] [Ganglia-developers]
Enumerating the
current value of metrics (eg load_one) on host graphics
To: Baker D.J. d.j.ba...@soton.ac.uk
Cc: ganglia-develop...@lists.sourceforge.net
ganglia-develop...@lists.sourceforge.net,
ganglia-general@lists.sourceforge.net
ganglia-general@lists.sourceforge.net
Message-ID:
alpine.DEB.2.00.101150250.20264@localhost6.localdomain6
Content-Type: text/plain; charset=utf-8
Reason why it's not showing is that legend is surpressed on
graphs of Size
small. You need to go to medium and above to see the legend.
Vladimir
On Thu, 10 Nov 2011, Baker D.J. wrote:
I have an interesting question for someone. I notice that on
older graphical interfaces that ? load_one? is
enumerated on each of the hosts in a cluster view/page. So
taking a look at our cluster pages we can get a quick
overview of the loads on each host. Using the new gweb I
don?t see the current value of ?load_one? displayed in
each of the host graphics. Some may say ?big deal?, however
this is very useful (as I say for a quick look at the
status of the hosts in the cluster). ?In the cluster view
that I have cited ? for example it would be useful to see
the current value of load_one on hosts like ?ff-grid3??.
--
Message: 2
Date: Mon, 14 Nov 2011 10:48:50 +0100
From: m.carpen m.car...@cineca.it
Re: [Ganglia-general] help for authorization in ganglia
That looks like your auth cookie isn't being set correctly when you log in. Offhand, I'm not sure why that should be. When you use login.php, is authentication successful, or do you see an error then? Do both of your users have this problem, or only one? If you're willing to check your cookies and email me the ganglia_auth cookie offlist, I could help further. alex On Nov 14, 2011, at 3:48 AM, m.carpen wrote: Hi all, I work in CINECA (Italy) and we've a Ganglia installation here. I've installed the last Web Front End: gweb v. 2.1.8. I created two users: hpc-dash and hpc-eni, we need now to add privileges. Following the twiki guide I've configured the apache server to enable authorization: http://sourceforge.net/apps/trac/ganglia/wiki/ganglia-web-2/AuthSystem users have been added and the ganglia secret has been set. When I login with user hpc-eni to login.php I find errors and the page cannot be correctly displayed, no graph is available. The error is showed on the web page: Warning: array_keys() [function.array-keys]: The first argument should be an array in /prod_meteo/web/hpc-dash/lib/GangliaAuth.php on line 28 Warning: Cannot modify header information - headers already sent by (output started at /prod_meteo/web/hpc-dash/lib/GangliaAuth.php:28) in /prod_meteo/web/hpc-dash/header.php on line 53 Do you have any idea? please I need help for that. thanks, Best Regards, Michele -- Michele Carpené SuperComputing Applications and Innovation Department CINECA - via Magnanelli, 6/3, 40033 Casalecchio di Reno (Bologna) - ITALY Tel: +39 051 6171xxx Fax: +39 051 6132198 http://hpc.cineca.it -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general -- RSA(R) Conference 2012 Save $700 by Nov 18 Register now http://p.sf.net/sfu/rsa-sfdev2dev1 ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] gweb 1.9.9.2607M - You do not have access to edit views.
On Oct 27, 2011, at 11:32 AM, Gerhard Lazu wrote: When I try to add an aggregate graph to a view I get: You do not have access to edit views. That's a message from the authentication system. http://sourceforge.net/apps/trac/ganglia/wiki/ganglia-web-2/AuthSystem As shipped, the gweb application is in read-only mode. You can either configure an authentication mechanism or disable authentication. alex -- The demand for IT networking professionals continues to grow, and the demand for specialized networking skills is growing even more rapidly. Take a complimentary Learning@Cisco Self-Assessment and learn about Cisco certifications, training, and career opportunities. http://p.sf.net/sfu/cisco-dev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Ganglia-general Digest, Vol 65, Issue 17
On Oct 27, 2011, at 7:03 PM, Daniel Burbano wrote: At the end, I started gmetad and get ganglia web page without graphs, just only information text of the server. While logged into trini2.uniandes.edu.co: 'telnet localhost 8649' will show you what gmond knows 'telnet localhost 8651' will show you what gmetad knows If gmond has no data, try starting it with a higher logging verbosity (globals debug_level = 10 in gmond.conf). Try your system logs to see if gmetad is having trouble writing RRD files. Maybe permissions or path issues. Try your Apache error logs to check for PHP (gweb) errors. Maybe the path to the rrdtool binary is incorrect. alex -- The demand for IT networking professionals continues to grow, and the demand for specialized networking skills is growing even more rapidly. Take a complimentary Learning@Cisco Self-Assessment and learn about Cisco certifications, training, and career opportunities. http://p.sf.net/sfu/cisco-dev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] gmetad dead but subsys locked
On Oct 25, 2011, at 8:08 AM, Daniel Burbano wrote: Hello, I installed ganglia-3.1.7 with the following steps: ./configure --sysconfdir=/etc/ganglia --with-librrd=/usr/local/rrdtool-1.3.1 --with-gmetad make make install Then I copied the directory /opt/ganglia/web inside of /var/www/html and I copied the gmetad.init file in /etc/init.d/. When I start the daemon, I got the following error: [root@trini2 ~]# service gmetad start Starting GANGLIA gmetad: [ OK ] [root@trini2 ~]# service gmetad status gmetad dead but subsys locked The httpd is running before to start gmetad. Any idea? You can run gmetad without the web code, so that's an unrelated issue. Try 'tail -f /var/log/messages' while your attempt to start gmetad. It should tell you what the problem is. alex -- The demand for IT networking professionals continues to grow, and the demand for specialized networking skills is growing even more rapidly. Take a complimentary Learning@Cisco Self-Assessment and learn about Cisco certifications, training, and career opportunities. http://p.sf.net/sfu/cisco-dev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] No folder for some hosts in /var/lib/ganglia/rrds/source, no graph on the web
Hi. Sounds like you solved this, but just in case:
- gmond does not create rrd files. it's in-memory only.
- gmetad will create rrd files
- if you change the way your sending metrics (udp_send_channel), you'll need
to change the way you're receiving metrics as well (udp_recv_channel).
- 'telnet localhost 8649' will tell you what gmond knows. again, that's
in-memory only.
- 'telnet localhost 8651' will tell you what gmetad knows. The data there
should correlate with what's in your RRDs.
alex
On Oct 23, 2011, at 11:00 PM, quanta wrote:
Hi,
gmond node version: 3.1.7-3
gmond head version: 3.0.7-1
ganglia-web-3.1.7-1
ganglia-3.0.7-1
ganglia-gmetad-3.0.7-1
It seems that this gmond node cannot create folder in
/var/lib/ganglia/rrds/source.
And on the web, it shows only the link, no graph. The other hosts still show
the graph normally.
Below is the permission:
ls -la /var/lib/ganglia/rrds/source/
total 88
drwxr-xr-x 22 ganglia ganglia 4096 Oct 22 13:43 .
drwxr-xr-x 11 ganglia ganglia 4096 Aug 1 18:22 ..
drwxr-xr-x 2 ganglia root4096 Oct 19 09:46 192.168.3.118
drwxr-xr-x 2 ganglia root4096 Jun 11 12:18 192.168.6.47
drwxr-xr-x 2 ganglia root4096 Oct 4 15:24 192.168.6.97
...
The error log:
ERROR: opening '/var/lib/ganglia/rrds/source/js1_6.98/load_one.rrd': No such
file or directory
ERROR: opening '/var/lib/ganglia/rrds/source/js2_6.109/load_one.rrd': No such
file or directory
(the source folder doesn't exist)
I only change the send channel section to use unicast:
udp_send_channel {
host = 192.168.x.x
port = 8649
}
On the gmond head, I can telnet to port 8649 on gmond node:
...
GANGLIA_XML VERSION=3.1.7 SOURCE=gmond
CLUSTER NAME=source LOCALTIME=1319428750 OWNER=unspecified
LATLONG=unspecified URL=unspecified
/CLUSTER
/GANGLIA_XML
What I tried:
stop gmetad
stop gmond head
stop gmond node
start gmond node
start gmond head
start gmetad
I enclosed here the screenshot.
What might be the culprit?
gmond - no
graph.png--
The demand for IT networking professionals continues to grow, and the
demand for specialized networking skills is growing even more rapidly.
Take a complimentary Learning@Cisco Self-Assessment and learn
about Cisco certifications, training, and career opportunities.
http://p.sf.net/sfu/cisco-dev2dev___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
--
The demand for IT networking professionals continues to grow, and the
demand for specialized networking skills is growing even more rapidly.
Take a complimentary Learning@Cisco Self-Assessment and learn
about Cisco certifications, training, and career opportunities.
http://p.sf.net/sfu/cisco-dev2dev
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] No folder for some hosts in /var/lib/ganglia/rrds/source, no graph on the web
One other thing: Looks like your error message says that the web frontend was
trying to retrieve RRD data for nodes which didn't exist (js1_6.98 and
js2_6.109). This probably means these were nodes which appeared in gmetad's
XML but which didn't have RRDs on the filesystem. Was this due to your
configuration changes?
I'm guessing your restart of gmetad gmond was the reason this stopped
happening, not the version upgrade itself, since the restart would clear the
gmetad XML which the web frontend retrieves. Just a theory.
On Oct 24, 2011, at 8:19 AM, Alex Dean wrote:
Hi. Sounds like you solved this, but just in case:
- gmond does not create rrd files. it's in-memory only.
- gmetad will create rrd files
- if you change the way your sending metrics (udp_send_channel), you'll need
to change the way you're receiving metrics as well (udp_recv_channel).
- 'telnet localhost 8649' will tell you what gmond knows. again, that's
in-memory only.
- 'telnet localhost 8651' will tell you what gmetad knows. The data there
should correlate with what's in your RRDs.
alex
On Oct 23, 2011, at 11:00 PM, quanta wrote:
Hi,
gmond node version: 3.1.7-3
gmond head version: 3.0.7-1
ganglia-web-3.1.7-1
ganglia-3.0.7-1
ganglia-gmetad-3.0.7-1
It seems that this gmond node cannot create folder in
/var/lib/ganglia/rrds/source.
And on the web, it shows only the link, no graph. The other hosts still show
the graph normally.
Below is the permission:
ls -la /var/lib/ganglia/rrds/source/
total 88
drwxr-xr-x 22 ganglia ganglia 4096 Oct 22 13:43 .
drwxr-xr-x 11 ganglia ganglia 4096 Aug 1 18:22 ..
drwxr-xr-x 2 ganglia root4096 Oct 19 09:46 192.168.3.118
drwxr-xr-x 2 ganglia root4096 Jun 11 12:18 192.168.6.47
drwxr-xr-x 2 ganglia root4096 Oct 4 15:24 192.168.6.97
...
The error log:
ERROR: opening '/var/lib/ganglia/rrds/source/js1_6.98/load_one.rrd': No such
file or directory
ERROR: opening '/var/lib/ganglia/rrds/source/js2_6.109/load_one.rrd': No
such file or directory
(the source folder doesn't exist)
I only change the send channel section to use unicast:
udp_send_channel {
host = 192.168.x.x
port = 8649
}
On the gmond head, I can telnet to port 8649 on gmond node:
...
GANGLIA_XML VERSION=3.1.7 SOURCE=gmond
CLUSTER NAME=source LOCALTIME=1319428750 OWNER=unspecified
LATLONG=unspecified URL=unspecified
/CLUSTER
/GANGLIA_XML
What I tried:
stop gmetad
stop gmond head
stop gmond node
start gmond node
start gmond head
start gmetad
I enclosed here the screenshot.
What might be the culprit?
gmond - no
graph.png--
The demand for IT networking professionals continues to grow, and the
demand for specialized networking skills is growing even more rapidly.
Take a complimentary Learning@Cisco Self-Assessment and learn
about Cisco certifications, training, and career opportunities.
http://p.sf.net/sfu/cisco-dev2dev___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
--
The demand for IT networking professionals continues to grow, and the
demand for specialized networking skills is growing even more rapidly.
Take a complimentary Learning@Cisco Self-Assessment and learn
about Cisco certifications, training, and career opportunities.
http://p.sf.net/sfu/cisco-dev2dev
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
--
The demand for IT networking professionals continues to grow, and the
demand for specialized networking skills is growing even more rapidly.
Take a complimentary Learning@Cisco Self-Assessment and learn
about Cisco certifications, training, and career opportunities.
http://p.sf.net/sfu/cisco-dev2dev
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Ganglia 3.2.0 and gweb 2.1.8
On Oct 24, 2011, at 9:14 PM, Jeff Blaine wrote: Hi all. Host is RHEL 5.7 x86_64. I'm reporting quite a few things in this message, as well as asking questions. I would appreciate any help. The gweb 2.1.8 README (which, sorry, is godawful formatted) says: This is an attempt to make the Ganglia UI more usable. You should be able to drop this into your $HTTPD_ROOT/ganglia then run make to generate conf_default.php. make install will create all the Aside from the sentence that ends abruptly (I didn't truncate it above, it actually ends like that), there's this, so perhaps the README needs updating/editing? html:rcf-monitor# make make: Nothing to be done for `default'. html:rcf-monitor# try 'make conf_default.php' The README also states this: You will need PHP JSON extension. It comes with PHP 5.2+. If you are on 5.1 use pecl install json pecl is part of php-pear package on RHEL/CentOS. Make sure you have # cat /etc/php.d/json.ini extension=json.so ... html:rcf-monitor# pecl install json No releases available for package pecl.php.net/json Cannot initialize 'json', invalid or missing package file Package json is not valid install failed html:rcf-monitor# Indeed, searching pear.php.net for 'json' returns no exact matches. Can anyone tell me the proper JSON module name that Ganglia Web would like as well as where to get it? Which version of PHP do you have? JSON support has been implemented in several ways in PHP, and gweb tries to use any which are available. 1. native JSON support or PECL JSON module (written in C, either in core PHP or added as a .so file later) 2. Services_JSON PEAR module (written in PHP, installed via PEAR) Run 'php -i | grep json'. If you see 'json support = enabled', then you're done. Otherwise see http://pecl.php.net/package/json. If that's not an option, look at installing http://pear.php.net/package/Services_JSON/. You can probably get away with just putting this single file somewhere in your include_path: http://svn.php.net/viewvc/pear/packages/Services_JSON/trunk/JSON.php?view=markup. http://us.php.net/manual/en/ini.core.php#ini.include-path /var/log/httpd/error_log shows this: [Mon Oct 24 21:35:25 2011] [error] [client 172.31.58.193] PHP Warning: include_once() [a href='function.include'function.include/a]: Failed opening '/var/www/html/lib/functions.php' for inclusion (include_path='/var/www/html/lib:.:/usr/share/pear:/usr/share/php') in /var/www/html/functions.php on line 10, referer: http://rcf-monitor.our.org/?m=load_oner=hours=by%20namehc=4mc=2 Any advice would be great. There's no ./lib/functions.php yet './functions.php' is trying to load one. Does line 10 of your functions.php look like this? https://github.com/vvuksan/ganglia-misc/blob/master/ganglia-web/functions.php#L10 That line is loading the wrapper code to use the PEAR JSON module if the native version isn't available. But it's not checking if the PEAR JSON module is actually available. If it's not, we should probably blow up with an error message at this point. Still, I'm not sure how that would produce the error message your reporting so there may be more going on. /var/log/httpd/error_log also shows this: ERROR: This RRD was created on other architecture ERROR: This RRD was created on other architecture This is a brand new install. There were no existing RRDs, so I don't see how that's possible. What gives? Any ideas? Are there any RRD files? That message isn't familiar, but I wonder if it's not finding any RRDs at all, and producing this message instead of something more helpful. If you have no conf_default.php, there's no way for gweb to actually find the RRD directories, so I wouldn't spend too much time on this error until you sort out the ones you've reported earlier. And finally, gweb is showing no generated images, yet I see nowhere to determine why that is. I'm not a PHP programmer. This is just a symptom of your other problems. You'll see images once they are sorted out. hth, alex -- The demand for IT networking professionals continues to grow, and the demand for specialized networking skills is growing even more rapidly. Take a complimentary Learning@Cisco Self-Assessment and learn about Cisco certifications, training, and career opportunities. http://p.sf.net/sfu/cisco-dev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Gmond Python NVIDIA GPU module questions.
If you can post up your graph.d file the names of the metrics you're trying to graph, we can help troubleshoot. On Oct 17, 2011, at 8:54 AM, Lee, Wayne wrote: To list, I’ve managed to get Ganglia 3.1.7 running with the Gmond Python NVIDIA GPU module and the output looks fine for 4 GPUs. However we have some 8 way GPU systems and want to display all 8 GPUs for these systems in the host view. After modifying and adding some extra files to the /var/www/html/ganglia/graph.d directory, I am able to see 8 GPUs. The problem that I am seeing is that the color graphs to the right of the metrics in host view don’t show up for all 8 GPUs. I only see the GPU memory graphs for the first four GPUs, 0-3.On our 4 GPU systems, I see all GPU related graphs which include the GPU memory used and GPU memory totals for all 4 GPUs as well as the graphs showing the total GPU and GPU memory utilizations for all four GPUs. I’m not much of a php web programmer, but does anyone know how I can correct this? Also, I do notice that regardless if I have 8 or 4 GPUs, I can’t seem to click on the color graphs for the GPUs in order to get a full size version to view. Does anyone know how to make this work with the GPU graphs. Again, I’m only having a problem displaying some of the color graphs. The individual black and white metric graphs appear to be fine. I wanted to attach a screenshot of what I am seeing, but I don’t think this will my posting to get posted to the mailing list. Any information would be greatly appreciated. Regards, Wayne Lee *** UNIX _IS_ user friendly, its just picky about who its friends are. *** This e-mail and any attachments are for the sole use of the intended recipient(s) and may contain information that is confidential. If you are not the intended recipient(s) and have received this e-mail in error, please immediately notify the sender by return e-mail and delete this e-mail from your computer. Any distribution, disclosure or the taking of any other action by anyone other than the intended recipient(s) is strictly prohibited. -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2d-oct___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2d-oct ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Ganglia Cluster aggregated graphs
If you're talking about aggregating some data from all the hosts in a cluster, take a look at the graphs in the graph.d/ directory. It's pretty easy to create your own. Example: https://github.com/vvuksan/ganglia-misc/blob/master/ganglia-web/graph.d/load_all_report.json This will graph the one/five/fifteen minute load averages for a cluster. alex On Oct 13, 2011, at 1:21 PM, Seth Graham wrote: On Oct 13, 2011, at 11:52 AM, Aidan Wong wrote: This is my first post on this Ganglia list =). I'm using the new Ganglia web 2.1.8 . Has anyone been able to create a graph that aggregates one common metric for several hosts. Try looking at the aggregate graphs tab on the web interface. It lets you use regular expressions to set up a graph showing many hosts at once. These graphs can also be added to views. -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2d-oct ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general -- All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense. http://p.sf.net/sfu/splunk-d2d-oct ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Setup of Multiple Grids. (Comments/Suggestions?)
A grid is the collection of all clusters known to 1 gmetad. I think what you'll need to do is to set up 1 gmetad server at each of your locations. us_tech_1, us_tech_2, etc. Somewhere in the world you'l have an additional gmetad which will aggregate data from all the other gmetads. alex On Sep 16, 2011, at 8:05 AM, Lee, Wayne wrote: To Ganglia List, I have some questions regarding Grids and Clusters with regards to Ganglia. First let me see if I understand what a “Grid” and “Cluster” are in relation to Ganglia. Cluster – Is a collection of computers (i.e. nodes) you wish to gather metrics on. From the Ganglia user interface, you can see graphs showing the aggregation of the metrics of all of the nodes you are gathering metrics for. Grid – Is a collection of clusters. From the Ganglia user interface, you can see graphs showing the aggregation of the metrics of all of clusters. Assuming I have the definitions correct, I would like to monitor and gather metrics for the following configuration. This configuration would monitor different clusters within different locations on a global scale. As an example, I work for a company with multiple offices globally where each office may have multiple clusters of computers to gather metrics for. Given what I know about Ganglia, can I configure a central Ganglia server to monitor the following setup? I would prefer to have a single Ganglia server which would collect all the information from the setup below and it would also be the web server which would host the Ganglia graphs.Is this possible and what would be the best configuration I should pursue?The arrangement below is hypothetical, but given that my employer will have different managers for the various groups of computers below in different countries, I figure that the setup below provide the managers with the information they would want. MyCompany_US_Tech1_Grid US_Tech1_Cluster_A US_Tech1_Cluster_B MyCompany_US_Tech2_Grid US_Tech2_Cluster_A US_Tech2_Cluster_B MyCompany_Europe_Tech1_Grid Europe_Tech1_Cluster_A Europe_Tech1_Cluster_B MyCompany_Europe_Tech2_Grid Europe_Tech2_Cluster_A Europe_Tech2_Cluster_B MyCompany_Asia_Tech1_Grid Asia_Tech1_Cluster_A Asia_Tech1_Cluster_B Any comments would be greatly appreciated? Wayne Lee *** UNIX _IS_ user friendly, its just picky about who its friends are. *** This e-mail and any attachments are for the sole use of the intended recipient(s) and may contain information that is confidential. If you are not the intended recipient(s) and have received this e-mail in error, please immediately notify the sender by return e-mail and delete this e-mail from your computer. Any distribution, disclosure or the taking of any other action by anyone other than the intended recipient(s) is strictly prohibited. -- BlackBerryreg; DevCon Americas, Oct. 18-20, San Francisco, CA http://p.sf.net/sfu/rim-devcon-copy2___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general -- BlackBerryreg; DevCon Americas, Oct. 18-20, San Francisco, CA http://p.sf.net/sfu/rim-devcon-copy2 ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Ganglia XML error: required
On Aug 29, 2011, at 4:23 AM, S 2011 wrote: Hi, In the ganglia web interface I do often see the XML errors like , There was an error collecting ganglia data (127.0.0.1:8656): XML error: required at 7776 There was an error collecting ganglia data (127.0.0.1:8656): XML error: required at 14268 There was an error collecting ganglia data (127.0.0.1:8656): XML error: required at 9872 how to track this error ? telnet localhost 8658 ganglia-output.xml xmllint ganglia-output.xml That, plus some visual inspection of the XML, should give you an idea what's wrong. alex -- EMC VNX: the world's simplest storage, starting under $10K The only unified storage solution that offers unified management Up to 160% more powerful than alternatives and 25% more efficient. Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Cluster setup
Have you configured any udp_recv_channel sections?
http://linux.die.net/man/5/gmond.conf
http://sourceforge.net/apps/trac/ganglia/wiki/Gmond%203.1.x%20General%20Configuration
On Aug 29, 2011, at 8:23 AM, Peter Ellevseth wrote:
Hi
I am quite new to ganglia, so I am having trouble setting it up like I want
it.
We are running a hpc-cluster and I want ganglia to monirot three sub-clusters.
Admin
Workstation
Computational
I have found some hints on how to do this, but nothing is working. I am able
to monitor a few machines individually, but not to collect them into
clusters. No I am doing
Gmetad.conf
Data_source “machine1” machine1
Data_source “machine2” machine2
Gmond.conf
Cluster {
Name = machine1
..
}
Udp_send_channel {
Host = machine1
Port = 8649
}
If I change host in send_channel to something else, and I do telnet to look
at the xml-file then no metrics are sent.
Anybody have any tips?
Regards Peter
--
EMC VNX: the world's simplest storage, starting under $10K
The only unified storage solution that offers unified management
Up to 160% more powerful than alternatives and 25% more efficient.
Guaranteed.
http://p.sf.net/sfu/emc-vnx-dev2dev___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
--
EMC VNX: the world's simplest storage, starting under $10K
The only unified storage solution that offers unified management
Up to 160% more powerful than alternatives and 25% more efficient.
Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] gweb2 RRDs directory is not readable
On Jul 29, 2011, at 9:06 PM, Bob Liu wrote:
Alex,
I thought the logic to include conf.php is already part of the
eval_conf.php... why do I have to include it again in conf_default.php?
Correct. My mistake.
My simple conf.php looks like the following:
?php
$conf['rrds'] = ${conf['gmetad_root']}/grid1/rrds; --- See my comment
below for this one... if I change it to /grid1/rrds instead of default
/rrds it would fail... I have tried it in conf_default.php as well...
does not work.
$conf['ganglia_port'] = 8672;
?
Can you be more specific about *how* it doesn't work? I wonder if a change
like this might make the problem a little clearer:
https://gist.github.com/0316c3ec956e2d96abb6
Just as a test, try defining $conf['rrds'] = '/var/lib/ganglia/grid1/rrds';,
without using a reference to $conf['gmetad_root'], and see if that helps.
And by the way... for some odd reason, I'm not able to move my rrds to any
other location except for the default /var/lib/ganglia/rrds same for
dwoo_compiled_dir, views_dir, conf_dir.. etc... it's expecting all
these to be under /var/lib/ganglia/dwoo and /var/lib/ganglia/conf... if I
change the location or name of the directory it would fail. It seems like
they are been hard-coded somewhere...
I don't find any hard-coded '/var/lib' references in the PHP code at the
moment. FWIW, I'm using an alternate RRDs directory and it's working OK.
alex
ps - I'm going to be away from email for a few days starting today, so
apologies if I'm slow to respond after this.
--
Got Input? Slashdot Needs You.
Take our quick survey online. Come on, we don't ask for help often.
Plus, you'll get a chance to win $100 to spend on ThinkGeek.
http://p.sf.net/sfu/slashdot-survey
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] gweb2 RRDs directory is not readable
On Jul 29, 2011, at 7:50 PM, Bob Liu wrote: I figured out the problem... for some reason it's not honoring my conf.php at all... so once I made all the changes directly in conf_default.php then the page finally came up after that... Does your conf_default.php include conf.php? That's how it's intended to work. The idea is that ganglia 'owns' conf_default.php and may replace or update values there in new releases. Any changes you need to make should be in conf.php. What's in your conf.php? If you could post it somewhere, we could help figure out the issue. alex -- Got Input? Slashdot Needs You. Take our quick survey online. Come on, we don't ask for help often. Plus, you'll get a chance to win $100 to spend on ThinkGeek. http://p.sf.net/sfu/slashdot-survey ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Rename gmetric group
How are you reporting your metrics? The gmetric protocol allows you to supply a 'group', but the gmetric binary itself doesn't (yet). Using gmetric library should allow you to send a group with your metrics. Otherwise you'll have to patch gmetric itself. http://tobym.posterous.com/gmetric-track-and-group-arbitrary-metrics-wit http://bugzilla.ganglia.info/cgi-bin/bugzilla/show_bug.cgi?id=295 http://www.mail-archive.com/ganglia-general@lists.sourceforge.net/msg06561.html http://code.google.com/p/embeddedgmetric/ https://github.com/igrigorik/gmetric alex On Jun 29, 2011, at 1:25 PM, grace rante wrote: Is there a way to provide a custom group name for Gmetric data graphs? Right now they are showing as No_group metrics and I'd like to change it if possible. Using GWEB2.0, Gmetad/Gmond 3.1.7, rrdtool 1.4.4 thanks, -- All of the data generated in your IT infrastructure is seriously valuable. Why? It contains a definitive record of application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-d2d-c2 ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general -- All of the data generated in your IT infrastructure is seriously valuable. Why? It contains a definitive record of application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-d2d-c2 ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] [Ganglia-developers] Announcing Ganglia Web 2.0RC1
On Jun 22, 2011, at 2:51 PM, Seth Graham wrote: On Jun 9, 2011, at 12:10 PM, Alex Dean wrote: I started off intending to allow per-view edit access, just like we allow per-cluster edit access for optional graphs. The complication is that each resource (a view or a cluster) in the ACL is only identified by a simple string. Thus you can't have a cluster and a view which share the same name - or, if you did you'd probably unwittingly be granting permissions you didn't mean to. I thought about introducing some kind of namespacing, and then just decided to punt until it was actually needed. So... maybe that time is now? :) Something like this wouldn't be too hard to implement: $acl-allowView( 'username', 'view-name', GangliaAcl::EDIT ); $acl-allowCluster( 'username', 'cluster-name', GangliaAcl::EDIT ); Please suggest alternate APIs here. That's just my initial brainstorm. I finally got a chance to sit down and poke at this. The good news is it's easy to implement a permissions system for adding graphs to an existing view. My method was to edit GangliaAcl.php to add an 'EDIT_VIEW' resource, and use the add() function along with a clustername to give a user view editing privileges. After updating the checkAccess() calls where appropriate in host_view.php and views.php, a user can add graphs to their view. That requires that the view name match the cluster name, right? Could you post your changes somewhere so we could see what you did? More complicated is the creation of the views themselves. Because views can have names without any relation to ganglia clusters, the ACL system won't work. I guess one could put in a restriction that a user can only create views with the same name of clusters they have edit permissions for, but that would limit them to owning a single view per cluster. (as an aside, is it intended that once a view is created, it cannot be removed via the web interface?) Probably more of an oversight. Patches welcome! :) The more I look at it, the more inclined I am to leave the configuration as it is. Every idea I come up with limits the flexibility of the Views or requires more acl maintenance in conf.php. I still want to take a stab at this, I just haven't had the time. Help me understand your use case better. You want to allow some non-admin users to edit a single view, right? Is there a case for limiting the visibility of a view, or are we only concerned with who can change a view? More generally, what permissions do we need? - view a view - create a view - edit a view - delete a view I'd say sensible defaulte are that admins can do all of these things for all views, and anonymous users can view all views which haven't been specifically hidden. alex -- Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Data protection magic? Nope - It's vRanger. Get your free trial download today. http://p.sf.net/sfu/quest-sfdev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] C api to create ganglia metrics
On Jun 16, 2011, at 3:34 AM, Indranil C wrote: Hi, Is there a C API, using which I can create ganglia metrics? Basically I want to avoid calliing gmetric, from my C code, as this seems to be too time and resource consuming for a large number of calls on a regular basis. Any thoughts? You can create gmond modules using the C API. http://sourceforge.net/apps/trac/ganglia/wiki/ganglia_gmond_c_modules alex -- EditLive Enterprise is the world's most technically advanced content authoring tool. Experience the power of Track Changes, Inline Image Editing and ensure content is compliant with Accessibility Checking. http://p.sf.net/sfu/ephox-dev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] ganglia web: can't select my cluster
On Jun 15, 2011, at 4:24 AM, Daems Dirk wrote: Hi Rick, Sorry for the confusion. Correction: My gmetad port is 8651. This is the one that is referred to in the conf.php file. My gmond port is 8649. If your installation is using the default port assignments, 8651 is the gmetad non-interactive XML port. You should use the interactive port (typically 8652) in your conf.php file. alex -- EditLive Enterprise is the world's most technically advanced content authoring tool. Experience the power of Track Changes, Inline Image Editing and ensure content is compliant with Accessibility Checking. http://p.sf.net/sfu/ephox-dev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] [Ganglia-developers] Announcing Ganglia Web 2.0RC1
On Jun 9, 2011, at 11:02 AM, Seth Graham wrote: On Jun 8, 2011, at 8:25 PM, Alex Dean wrote: Hi Seth. I'm just back from a week off the grid, and trying to get caught up on a mountain of electronic stuff. Here's my quick response. Please let me know if more explanation is required. Nope, the explanation makes sense. The only thing I was missing was detail about the philosophy behind the privileges system. Editing views is not per-cluster permission because views can contain graphs from many clusters. Currently, we only support a single 'edit' permission for all views. (A user can either edit all views, or can edit none.) You can't selectively grant edit permission on a single view. That restriction could possible be lifted in the future if there is demand for it. It's my primary motivation for updating to the new interface, actually. Interesting. I started off intending to allow per-view edit access, just like we allow per-cluster edit access for optional graphs. The complication is that each resource (a view or a cluster) in the ACL is only identified by a simple string. Thus you can't have a cluster and a view which share the same name - or, if you did you'd probably unwittingly be granting permissions you didn't mean to. I thought about introducing some kind of namespacing, and then just decided to punt until it was actually needed. So... maybe that time is now? :) Something like this wouldn't be too hard to implement: $acl-allowView( 'username', 'view-name', GangliaAcl::EDIT ); $acl-allowCluster( 'username', 'cluster-name', GangliaAcl::EDIT ); Please suggest alternate APIs here. That's just my initial brainstorm. alex -- EditLive Enterprise is the world's most technically advanced content authoring tool. Experience the power of Track Changes, Inline Image Editing and ensure content is compliant with Accessibility Checking. http://p.sf.net/sfu/ephox-dev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] [Ganglia-developers] Announcing Ganglia Web 2.0RC1
Hi Seth. I'm just back from a week off the grid, and trying to get caught up
on a mountain of electronic stuff. Here's my quick response. Please let me
know if more explanation is required.
If you want to allow a user to be able to edit views, you need to grant them
edit permission on GangliaAcl::ALL_VIEWS.
You can allow users to edit:
- All views.
$acl-allow( 'username', GangliaAcl::ALL_VIEWS, GangliaAcl::EDIT );
- Optional graphs for all clusters.
$acl-allow( 'username', GangliaAcl::ALL_CLUSTERS, GangliaAcl::EDIT );
- Optional graphs for an individual cluster.
$acl-allow( 'username', 'clustername', GangliaAcl::EDIT );
Editing views is not per-cluster permission because views can contain graphs
from many clusters. Currently, we only support a single 'edit' permission for
all views. (A user can either edit all views, or can edit none.) You can't
selectively grant edit permission on a single view. That restriction could
possible be lifted in the future if there is demand for it.
Hope that helps. As I said, please let me know if further explanation is
required.
alex
On Jun 7, 2011, at 8:20 PM, Vladimir Vuksan wrote:
Sounds like a bug. We'll take a look. If you end up fixing it feel free to
send us a patch.
Thanks,
Vladimir
On Tue, 7 Jun 2011, Seth Graham wrote:
I'm having some issues getting the user roles working as expected.
The wiki instructs something like:
$acl-addRole( $username, GangliaAcl::GUEST );
$acl-allow( $username, $cluster, GangliaAcl::EDIT );
Which does not result in the little blue + sign to be drawn next to graphs.
From line 71 in host_view.php, there is this line:
if(checkAccess(GangliaAcl::ALL_VIEWS, GangliaAcl::EDIT, $conf)) {
Changing it to:
if(checkAccess($clustername, GangliaAcl::EDIT, $conf)) {
Allows the check to succeed, but I run into the same problem in views.php.
What does the 'EDIT' role actually allow a user to edit, if not views? And
is it possible to configure the interface to allow a user to only edit
specific views? As configured now, it appears view editing is all or nothing.
thanks,
On Jun 1, 2011, at 10:08 AM, Vladimir Vuksan wrote:
Announcing Ganglia Web 2.0 Release Candidate 1.
http://ganglia.info/?p=373
Vladimir
--
Simplify data backup and recovery for your virtual environment with vRanger.
Installation's a snap, and flexible recovery options mean your data is safe,
secure and there when you need it. Data protection magic?
Nope - It's vRanger. Get your free trial download today.
http://p.sf.net/sfu/quest-sfdev2dev
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
--
EditLive Enterprise is the world's most technically advanced content
authoring tool. Experience the power of Track Changes, Inline Image
Editing and ensure content is compliant with Accessibility Checking.
http://p.sf.net/sfu/ephox-dev2dev
___
Ganglia-developers mailing list
ganglia-develop...@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers
--
EditLive Enterprise is the world's most technically advanced content
authoring tool. Experience the power of Track Changes, Inline Image
Editing and ensure content is compliant with Accessibility Checking.
http://p.sf.net/sfu/ephox-dev2dev
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] new Web interface ETA?
On May 10, 2011, at 7:23 AM, Vladimir Vuksan wrote: One of the big features missing has been the authentication system which we deemed important as the web app is now writing to disk. Alex Dean (@mostlyalex on Twitter) has been making steady progress on that. I think that the authentication code is mostly finished. What's lacking (of course) is testing documentation. I'll have time to work on it again later this week. The plan right now is to ship the new frontend in read-only mode. To allow web-based editing of various features (optional graphs, views, etc) you'll need to configure authentication. Authentication will be provided by the web server, so you can use htpasswd, ldap, etc. This is easy to do in Apache, and I'm planning to work out instructions for Nginx as well since we've had at least one person report using it. Is anyone using lighttpd? I'm sure we can come up with some configuration instructions for it as well if desired. alex -- Achieve unprecedented app performance and reliability What every C/C++ and Fortran developer should know. Learn how Intel has extended the reach of its next-generation tools to help boost performance applications - inlcuding clusters. http://p.sf.net/sfu/intel-dev2devmay ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Retain the MAX values in RRD?
On Apr 28, 2011, at 8:28 AM, Indranil C wrote: However, RRD file stores only the average, so where do I catch the actual absolute value? Is it in gmond? Please help. Thanks. Look for RRAs config in gmetad.conf. Saving the average is gmetad's default behavior, but you can change this if you desire. I think this can only be specified at rrd creation time, though. This may be helpful: http://oss.oetiker.ch/rrdtool/doc/rrdcreate.en.html alex -- WhatsUp Gold - Download Free Network Management Software The most intuitive, comprehensive, and cost-effective network management toolset available today. Delivers lowest initial acquisition cost and overall TCO of any competing solution. http://p.sf.net/sfu/whatsupgold-sd ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
[Ganglia-general] who is using 'private clusters' feature?
Hi all. We're working to provide some access controls for the new features in the upcoming 2.0 release of Ganglia Web. This will largely replace the existing 'private clusters' functionality. I'd like to hear from folks who are currently using this feature, to ensure that our replacement continues to meet your needs. Currently: - You may define clusters as private, and configure a password which is used by anyone who wants to view that cluster. - There is a 'controlroom' context which grants access to all clusters to anyone who knows the password. Upcoming: - Private clusters are still hidden from guests. - Admin users can view edit all clusters (public private). I think this will replace the 'controlroom' context. - Apache will provide authentication. Every user who will have elevated privileges will need a username password. The private_clusters config file will be removed. A configuration file in Ganglia Web will determine which clusters are private, and which users may access them. Access rules are configurable, so I think just about any permission scheme can be implemented. Intermediate levels of access (between guest and admin) should be easy to implement. The big change is going from '1 password per cluster' to '1 password per user', and moving authentication from Ganglia to Apache. If you still want to have '1 password per cluster', you could just create a single 'user' for each of your private clusters, and share this amongst all the people who should be able to access the private cluster. Looking forward to your feedback! best, alex -- Benefiting from Server Virtualization: Beyond Initial Workload Consolidation -- Increasing the use of server virtualization is a top priority.Virtualization can reduce costs, simplify management, and improve application availability and disaster protection. Learn more about boosting the value of server virtualization. http://p.sf.net/sfu/vmware-sfdev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] How to plot metrics without using gmetric
On Apr 1, 2011, at 10:44 AM, Vladimir Vuksan wrote: You have to do the work then. I have a script that calculates queries per second by getting the count of queries from this sampling interval subtracting it from a previous interval then dividing it by number of seconds passed and voila you have queries/sec. You then use gmetric to send that to Ganglia. I've implemented something like this myself, and I've always thought it was a hassle. I noticed that rrdcreate seems to allow creation of rrd file which accept an absolute value but actually store a rate of change (derived from current value, previous value, and time delta). http://oss.oetiker.ch/rrdtool/doc/rrdcreate.en.html Look for the 'DST' section, and the 'COUNTER' type. Wonder if we could add some configuration syntax to gmetad.conf to allow per-metric rrd creation options, and use COUNTER for a rate metric? I see 'support for counters' at http://sourceforge.net/apps/trac/ganglia/wiki/ganglia_wish-list. Is this the same thing? alex -- Create and publish websites with WebMatrix Use the most popular FREE web apps or write code yourself; WebMatrix provides all the features you need to develop and publish your website. http://p.sf.net/sfu/ms-webmatrix-sf ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Network bytes spikes
On Mar 30, 2011, at 12:42 PM, Cameron L. Spitzer wrote: As I said, it's a known bug that will never be fixed by ganglia's maintainers, so if you want it fixed you have to try to do it yourself. The bug is missing input sanity checking. If you're willing/able to fix this (or you already have a patch?) it would be great to have the fix submitted back to Ganglia as well. alex -- Create and publish websites with WebMatrix Use the most popular FREE web apps or write code yourself; WebMatrix provides all the features you need to develop and publish your website. http://p.sf.net/sfu/ms-webmatrix-sf ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Including a Head Node in Ganglia Monitoring
On Mar 9, 2011, at 12:40 PM, Jeffrey L Moon wrote:
36 udp_send_channel {
43 host = head-eth0
44 port = 8649
45 ttl = 1
46 }
Does anybody know what I might be doing wrong that prevents the compute nodes
from reporting their metrics?
You're only sending your metrics to the head node. If you want them available
via 'telnet localhost 8469' on the compute node, you need to define another
udp_send_channel to send the metrics to localhost.
The head node is receiving your metrics in this configuration, right?
alex
--
Colocation vs. Managed Hosting
A question and answer guide to determining the best fit
for your organization - today and in the future.
http://p.sf.net/sfu/internap-sfd2d
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Including a Head Node in Ganglia Monitoring
Do you want the head node to be reported as part of the same cluster as the
compute nodes, or as a separate cluster? I think you're saying you want the
head node to report as part of the same cluster, in which case you just need to
add a udp_send_channel to the head node's gmond.conf.
So, head is reporting to itself and will show up just like any other cluster
node when gmetad polls for data.
upd_send_channel {
host = head
port = 8649
}
alex
On Mar 3, 2011, at 2:08 PM, Jeffrey L Moon wrote:
I am trying to configure ganglia to monitor a cluster. My initial thoughts
were to have a gmond running on each of the compute nodes,
and to have a gmond running on the head node which would aggregate the data
from the compute nodes and add it's own monitoring data.
The problem I am having is that it kind of appears that I have to choose to
either aggregate the data at the head node, or monitor the head node,
but I can't seem to get both working at the same time.
I am using the unicast option.
On the compute nodes:
cluster {
name = my cluster
}
host {
location = unspecified
}
udp_send_channel {
host = head
port = 8649
}
tcp_accept_channel {
port = 8649
}
On the head node:
cluster {
name = my cluster
}
host {
location = unspecified
}
udp_recv_channel {
port = 8649
}
tcp_accept_channel {
port = 8649
}
Then I have a gmetad running on a seperate server with the following in it's
gmetad.conf:
data_source my cluster head
Any help or suggestions would be greatly appreciated!
Jeff
--
Free Software Download: Index, Search Analyze Logs and other IT data in
Real-Time with Splunk. Collect, index and harness all the fast moving IT data
generated by your applications, servers and devices whether physical, virtual
or in the cloud. Deliver compliance at lower cost and gain new business
insights. http://p.sf.net/sfu/splunk-dev2dev
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
--
What You Don't Know About Data Connectivity CAN Hurt You
This paper provides an overview of data connectivity, details
its effect on application quality, and explores various alternative
solutions. http://p.sf.net/sfu/progress-d2d
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Web Interface not showing Clusters / hosts
There are a few places you can verify that everything is matching up. Everything I'm writing here is for unicast. I'm not sure how much of the gmond info applies to multicast, but the gmetad web parts should be the same for uni/multicast. Unicast clusters have normal nodes head nodes. A head node is one which knows the state of all other nodes in the cluster, and can be polled by gmetad. Just for redundancy and clarity and redundancy: A head node is also a 'normal' node, in that it knows it's own state as well as others'. 1. normal gmond udp_send_channel (hostport) must match a head-node udp_recv_channel. 2. gmetad's data_source must match a head-node gmond tcp_accept_channel. 3. web's conf.php $ganglia_ip/$ganglia_port must match gmetad's xml_port. Your XML looks like 1 2 are true, but I'd still double-check. What about #3? I've never seen the symptoms you're describing, so I'm guessing to some degree. You might try shutting down all the daemons, and bringing up just 1 gmond gmetad. See how that looks, then bring up more gmonds one at a time, and see where things start to fail. Check gmond gmetad XML output at every stage, as well as the web. alex On Dec 21, 2010, at 3:52 PM, Tarabas wrote: Hi Bernard, I am setting it up with unicast. The structure is as follwows: server-A - Port 8649 gmond/gmetad server-B - Port 8661 gmond Both clusters just have only one host (same server). Gmetad on Server-A collects data from Server-A and Server-B gmond. data_source server-A ip-server-A:8649 data_source server-B ip-server-B:8660 The XML from the gmetad looks like this, which in my view looks okay: [...] GANGLIA_XML VERSION=3.1.7 SOURCE=gmetad GRID NAME=mediaskill AUTHORITY=http://smurfette/ganglia/; LOCALTIME=1292967617 CLUSTER NAME=server-A LOCALTIME=1292967606 OWNER=mediaskill LATLONG=unspecified URL=unspecified HOST NAME=smurfette IP=ip-smurfette REPORTED=1292967599 TN=18 TMAX=20 DMAX=0 LOCATION=Berlin GMOND_STARTED=1292965419 [...] /HOST /CLUSTER CLUSTER NAME=server-B LOCALTIME=1292967609 OWNER=mediaskill LATLONG=unspecified URL=unspecified HOST NAME=eva IP=ip-eva REPORTED=1292967606 TN=11 TMAX=20 DMAX=0 LOCATION=Berlin GMOND_STARTED=1292963565 [...] /HOST /CLUSTER /GRID /GANGLIA_XML The web-frontend always defaults to Server-A and is not able to correctly display any other server (B, C, D) which I also added in the same manner with increasing port numbers starting at 8660. I did not see any errors with debug enabled in any of the gmond or the gmetad ... only the web interface seems to have some problems displaying the hosts. I configured it to the 8651 port of the gmetad on localhost. Best regards Manuel Tuesday, December 21, 2010, 10:10:08 PM, you wrote: BL Hi Manuel: BL Can you please clarify whether you are trying to setup Ganglia with BL unicast or multicast? BL You can also get more troubleshooting information by running BL gmetad/gmond in debug mode (-d 2) and looking at your apache error BL logs. BL Cheers, BL Bernard -- Forrester recently released a report on the Return on Investment (ROI) of Google Apps. They found a 300% ROI, 38%-56% cost savings, and break-even within 7 months. Over 3 million businesses have gone Google with Google Apps: an online email calendar, and document program that's accessible from your browser. Read the Forrester report: http://p.sf.net/sfu/googleapps-sfnew ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general -- Forrester recently released a report on the Return on Investment (ROI) of Google Apps. They found a 300% ROI, 38%-56% cost savings, and break-even within 7 months. Over 3 million businesses have gone Google with Google Apps: an online email calendar, and document program that's accessible from your browser. Read the Forrester report: http://p.sf.net/sfu/googleapps-sfnew ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Writing metrics
On Dec 6, 2010, at 3:16 PM, Antonio Óscar Balmaseda wrote:
Hi, everyone,
I have a strange problem. I'm writing a new metric for ganglia, in python, in
order to measure some data of the apache log. I'm pretty sure that the code
is correct but the thing is that it's absolutely impossible open any file.
When I try do it, the system returns:
Traceback (most recent call last):
File /usr/lib/ganglia/python_modules/test.py, line 41, in Apache_Response
fileHand = open('/var/log/apache2/access.log','r')
IOError: [Errno 13] Permission denied: '/var/log/apache2/access.log'
I've checked the permissions, tested with other files in other directories
and done everything that I could think. Anyone has a clue?
Are you sure you've checked as the gmond user (nobody,etc) in addition to your
normal user account ?
alex
--
What happens now with your Lotus Notes apps - do you make another costly
upgrade, or settle for being marooned without product support? Time to move
off Lotus Notes and onto the cloud with Force.com, apps are easier to build,
use, and manage than apps on traditional platforms. Sign up for the Lotus
Notes Migration Kit to learn more. http://p.sf.net/sfu/salesforce-d2d
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Plus sign in the metric name - graph not shown
On Sep 14, 2010, at 2:03 PM, Bernard Li wrote: Hi Urs: On Tue, Sep 14, 2010 at 11:29 AM, Urs Beyerle urs.beye...@env.ethz.ch wrote: I don't know if this bug is related to my problem, since + is allowed in filenames under Linux. But maybe php has a problem with the plus sign? I found that if I replace in the URL + with %2B it works. So if the ganglia php script (which?) would replace + with %2B it would work. Try looking at graph.php and friends -- could you please file a bug at bugzilla.ganglia.info as well? The web frontend actually provides a debug mode. Right click on an image (even if there is no image you can still right-click on the text where there is supposed to be an image) and select show image. Modify the URL and add debug=3, that will show you the rrdtool command that was used to generate that image -- it might help with debugging. The code might need calls to rawurlencode() when creating URLs, and rawurldecode() when reading URL arguments. echo( rawurlencode( '+' ) ); %2B echo( rawurldecode( '%2B' ) ); + http://www.php.net/manual/en/function.rawurlencode.php alex -- Start uncovering the many advantages of virtual appliances and start using them to simplify application deployment and accelerate your shift to cloud computing. http://p.sf.net/sfu/novell-sfdev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Help understanding gmond setup
On Aug 18, 2010, at 3:16 PM, Stevens, Weston J wrote: I'm guessing with this setup all the other gmonds in the cluster send their data via UDP unicast to the gmond located on the head node which acts like a hub, which connects via TCP to the gmetad located on the same machine and relays the data? gmetad initiates a TCP connection to gmond to get metric data, rather than the other way around. Other than that, I think your assessment of the situation seems correct. alex -- This SF.net email is sponsored by Make an app they can't live without Enter the BlackBerry Developer Challenge http://p.sf.net/sfu/RIM-dev2dev ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] Multiple gmonds or gmetads on same host
You could also write a single init script which reads which config
file to use from an environment variable. That would cut down on the
number of copy/paste scripts.
# export GMOND_CONF=/etc/ganglia/gmond.conf-ramdisk
# /etc/init.d/gmond start
Then /etc/init.d/gmond would do something like '/usr/sbin/gmond -c
$GMOND_CONF'. Not sure if that makes your task easier or not, but I
use an approach like this at my work for a similar purpose.
alex
On Aug 2, 2010, at 5:58 PM, Stevens, Weston J wrote:
MUCH obliged. I'll try this stuff out tomorrow and let you guys know
how things go.
-Original Message-
From: Jesse Becker [mailto:haw...@gmail.com]
Sent: Monday, August 02, 2010 3:19 PM
To: Rick Cobb
Cc: Stevens, Weston J; ganglia-general@lists.sourceforge.net
Subject: Re: [Ganglia-general] Multiple gmonds or gmetads on same host
Ah, then I mis-understood.
Yes, you will need different configurations for this, and will need
to make use of the -c option (and -p, if you want to track the PID
files). The different configuration files will still need to make
use of different ports though. Gmond will run quite happily from
the command line.
You can start gmond and gmetad without their respective startup
scripts--neither one does anything fancy, and both are largely
boilerplate scripts from Red Hat.
If you want to make a startup script for your test versions then you
will need to make a new file in /etc/init.d/ for each new instance
you want to run. So /etc/init.d/gmond_test-ramdisk and /etc/init.d/
gmond_test-3.1.7, for example.
As mentioned, you will need to use different configuration files for
each gmond instance. It should be simple to add to this to each of
the startup scripts. For example, on my Centos system, the gmond
script has a line at the top that reads:
GMOND=/usr/sbin/gmond
This can be changed to
GMOND=/usr/sbin/gmond -c /etc/ganglia/gmond.conf-ramdisk
and nothing else should need to change in the rest of the file--so a
minor change may be warranted, *if* you want to use init scripts
(the use of which is not actually required).
On Mon, Aug 2, 2010 at 18:07, Rick Cobb rick_c...@ieee.org wrote:
That's a different use-case than I understood Mr. Stevens to be
asking
about. As I understand it, he wants to test ganglia configurations
themselves, not just segment his grid between test clusters
development clusters.
E.g., he might want to model his clusters' memberships differently,
or
run on RRDcache instead of ramdisk, or find out if the latest Python
module locks up gmond, etc., without affecting his ability to measure
the stuff his team knows works.
In that case, he will need to use the '-c' option when starting his
daemons. OTOH, there's no reason to touch the stuff in
/etc/init.d/functions; *nobody* needs to touch those outside of the
fedora/redhat/etc., folks.
(That does remind me that you'll probably also need to use the '-p'
option to set the pidfile name of each daemon instance.)
-- ReC
On Mon, Aug 2, 2010 at 2:29 PM, Jesse Becker haw...@gmail.com
wrote:
No, I don't think so.
I have a server that collects metrics from three different ganglia
clusters, and just have multiple udp_recv_channel stanzas, one for
each unicast port, and another for each multicast port. I only have
gmond process running, bound to multiple ports. For example:
udp_recv_channel {
mcast_join = 239.2.11.71
port = 8649 /* cluster */
bind = 239.2.11.71
}
udp_recv_channel {
mcast_join = 239.2.11.71
port = 8648 /* workstations */
bind = 239.2.11.71
}
udp_recv_channel {
port = 8649 /* cluster */ } udp_recv_channel {
port = 8648 /* workstations*/ }
# lsof -i |grep gmond
gmond 3950 nobody3u IPv432629 UDP
239.2.11.71:8649 gmond 3950 nobody4u IPv432633
UDP 239.2.11.71:8648 gmond 3950 nobody6u IPv4
32635
UDP *:8648 gmond 3950 nobody7u IPv432636
UDP
*:8649 gmond 3950 nobody8u IPv432637 TCP
*:8649 (LISTEN)
Now, the gmond.conf files in the cluster and workstation groups
are different, since they need to send to differen port numbers.
On Mon, Aug 2, 2010 at 16:39, Stevens, Weston J
weston.j.stev...@boeing.com wrote:
Wouldn't I need different versions of gmetad.conf and gmond.conf,
like gmetad-test.conf and gmond-test.conf for instance, in order
to have different ganglia configurations running at the same
time? And wouldn't I need separate startup scripts for running
these different versions, in which case the test version startups
would need to be changed with the -c option?
-Original Message-
From: Jesse Becker [mailto:haw...@gmail.com]
Sent: Monday, August 02, 2010 12:56 PM
To: Stevens, Weston J
Cc: Rick Cobb; ganglia-general@lists.sourceforge.net
Re: [Ganglia-general] Multiple gmonds or gmetads on same host
On Aug 3, 2010, at 1:33 PM, Stevens, Weston J wrote: I've been able to duplicate gmond by changing the port for the tcp_accept_channel, however I have been unable to duplicate gmetad (how is this done in gmetad.conf exactly?). Each gmetad needs unique ports as well as a separate directory to write rrds to. you can start it with the same -c and -f arguments that gmond supports. 'gmetad -h' will give you more information. There should be an example gmetad.conf in the source code which shows what all the different configurable options are. I think you'll need to change xml_port, interactive_port, and rrd_rootdir. If an unprivileged user is starting the daemon, you'll probably also need to 'setuid off'. I also lack permission to create new scripts in /etc/init.d so I have them running them from elsewhere, I fear I will need to copy and edit /etc/init.d/functions in order to do this? functions assumes you'll write to directories likd /var/run, and you won't be able to do that as a non-root user. I'm sure you can make it work, but you'll have to hack it a bit. alex -- The Palm PDK Hot Apps Program offers developers who use the Plug-In Development Kit to bring their C/C++ apps to Palm for a share of $1 Million in cash or HP Products. Visit us here for more details: http://p.sf.net/sfu/dev2dev-palm ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] cluster nodes disagree on 1 machine's hostname
On Jul 30, 2010, at 11:27 AM, Rick Cobb wrote: It'll do a reverse lookup on the IP address the metric packet came from. Names in the configuration files are irrelevant; if, for example, your packet is routed on a different interface than you expect, the host will be named after whatever you've named that interface. gmond on dev-1-dist1 shows HOST NAME=10.0.3.31 rather than HOST NAME=dev-1-dist2.meteostar.local. When I log into dev-1-dist1 and do a reverse lookup on 10.0.3.31 as the user running gmond, I get 'dev-1-dist2.meteostar.local'. -bash-3.2$ hostname dev-1-dist1.meteostar.local -bash-3.2$ dig -x 10.0.3.31 +short dev-1-dist2.meteostar.local. If gmond should display the results of a reverse-lookup, I'm not sure why the IP address continues to appear in the gmond XML. These machines have only a single IP address ethernet interface (aside from the local loopback) so I don't think it's a question of packets traveling a different route. There are a number of email threads about this in the archives; the mcast_bind parameter can be helpful, as can making sure your hosts are routing the way you expect them to. Thanks. I've tried several searches and not found anything relevant. Do you recall subject lines or dates for the posts you were thinking of? The machines are using unicast not multicast. alex -- The Palm PDK Hot Apps Program offers developers who use the Plug-In Development Kit to bring their C/C++ apps to Palm for a share of $1 Million in cash or HP Products. Visit us here for more details: http://p.sf.net/sfu/dev2dev-palm ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
Re: [Ganglia-general] cluster nodes disagree on 1 machine's hostname
I ran through the series of uname hostname commands you suggest, and all appear to return correct values on both dev-1-dist1 and dev-1-dist2. I'm not sure how these commands are relevant, though. uname hostname will report the name of the local machine, but my problem seems to be gmond on dev-1-dist1 not reporting a hostname for dev-1- dist2, even though reverse DNS for dev-1-dist2's IP address appears to be set up correctly. If this has been addressed in the archives, could someone steer me towards some relevant posts? I've been looking but not found anything yet. thanks, alex On Aug 2, 2010, at 9:29 AM, Peter Senna Tschudin wrote: Hi Alex, Setting up hostname and domain is not a simple task because there is more than one way of doing it. To check your hostname and domain, you need to run: uname -n hostname -a hostname -s hostname -d hostname -f hostname If all these commands return the correct values, then the hostname and domain are configured correctly. Source: http://www.debianadmin.com/change-hostname-or-server-name-of-a-linux-machine.html Peter On Mon, Aug 2, 2010 at 11:19 AM, Alex Dean a...@crackpot.org wrote: On Jul 30, 2010, at 11:27 AM, Rick Cobb wrote: It'll do a reverse lookup on the IP address the metric packet came from. Names in the configuration files are irrelevant; if, for example, your packet is routed on a different interface than you expect, the host will be named after whatever you've named that interface. gmond on dev-1-dist1 shows HOST NAME=10.0.3.31 rather than HOST NAME=dev-1-dist2.meteostar.local. When I log into dev-1-dist1 and do a reverse lookup on 10.0.3.31 as the user running gmond, I get 'dev-1-dist2.meteostar.local'. -bash-3.2$ hostname dev-1-dist1.meteostar.local -bash-3.2$ dig -x 10.0.3.31 +short dev-1-dist2.meteostar.local. If gmond should display the results of a reverse-lookup, I'm not sure why the IP address continues to appear in the gmond XML. These machines have only a single IP address ethernet interface (aside from the local loopback) so I don't think it's a question of packets traveling a different route. There are a number of email threads about this in the archives; the mcast_bind parameter can be helpful, as can making sure your hosts are routing the way you expect them to. Thanks. I've tried several searches and not found anything relevant. Do you recall subject lines or dates for the posts you were thinking of? The machines are using unicast not multicast. alex -- The Palm PDK Hot Apps Program offers developers who use the Plug-In Development Kit to bring their C/C++ apps to Palm for a share of $1 Million in cash or HP Products. Visit us here for more details: http://p.sf.net/sfu/dev2dev-palm ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general -- Peter Senna Tschudin peter.se...@gmail.com gpg id: 48274C36 -- The Palm PDK Hot Apps Program offers developers who use the Plug-In Development Kit to bring their C/C++ apps to Palm for a share of $1 Million in cash or HP Products. Visit us here for more details: http://p.sf.net/sfu/dev2dev-palm ___ Ganglia-general mailing list Ganglia-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ganglia-general
[Ganglia-general] cluster nodes disagree on 1 machine's hostname
In one of my tiny (2 node) clusters, the gmond XML on 1 machines does
not report the correct hostname for the other machine. I'm trying to
figure out why.
How does gmond determine what goes in the HOST NAME= attribute?
My gmond.conf includes:
udp_send_channel {
host = dev-1-dist1
port = 8649
ttl = 1
}
udp_send_channel {
host = dev-1-dist2
port = 8649
ttl = 1
}
On dev-1-dist2:
$ telnet localhost 8649 | grep HOST
HOST NAME=dev-1-dist2.meteostar.local IP=10.0.3.31
REPORTED=1280503440 TN=16 TMAX=20 DMAX=0 LOCATION=
GMOND_STARTED=1280502260
HOST NAME=dev-1-dist1.meteostar.local IP=10.0.3.32
REPORTED=1280503454 TN=2 TMAX=20 DMAX=0 LOCATION=
GMOND_STARTED=1280502234
On dev-1-dist1:
$ telnet localhost 8649 | grep HOST
HOST NAME=10.0.3.31 IP=10.0.3.31 REPORTED=1280504565
TN=1 TMAX=20 DMAX=0 LOCATION=unspecified
GMOND_STARTED=1280504465
HOST NAME=dev-1-dist1.meteostar.local IP=10.0.3.32
REPORTED=1280504565 TN=1 TMAX=20 DMAX=0 LOCATION=
GMOND_STARTED=1280504550
Both machines use the same DNS servers, and there are correct reverse
lookup records for those IPs.
[r...@dev-1-dist1 software]# nslookup 10.0.3.32
Server: 10.0.0.5
Address: 10.0.0.5#53
32.3.0.10.in-addr.arpa name = dev-1-dist1.meteostar.local.
[r...@dev-1-dist1 software]# nslookup 10.0.3.31
Server: 10.0.0.5
Address: 10.0.0.5#53
31.3.0.10.in-addr.arpa name = dev-1-dist2.meteostar.local.
What else might I need to investigate?
thanks,
alex
--
The Palm PDK Hot Apps Program offers developers who use the
Plug-In Development Kit to bring their C/C++ apps to Palm for a share
of $1 Million in cash or HP Products. Visit us here for more details:
http://p.sf.net/sfu/dev2dev-palm
___
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
